Performance of the MCP-PMTs of the TOP counter in the first beam operation of the Belle II experiment K. Matsuoka (KMI, Nagoya Univ.) on behalf of the Belle II TOP group 5th International Workshop on New Photon-Detectors (PD18), Tokyo, Nov. 29, 2018
The Belle II experiment 2 B-factory experiments Next generation B-factory experiment BaBar Belle Belle II Confirmed Kobayashi-Maskawa theory with > 1 ab 1 data Search for new physics via precision measurements with 50 ab 1 data TOP Challenge on the detector Cope with harsh beam background Improve the performance Barrel PID TOP counter
TOP counter State-of-the-art Cherenkov ring imaging detector K/p identification by means of b reconstruction using precise timing measurement of internally reflected Cherenkov photons 3 K or p TOP 2 cm Mirror p K q C Air (n=1) Quartz (n=1.47) @ 400 nm TOF (~1 m) 1.5 T Air (n=1) Time of propagation (TOP) e + cos θ C = 1 e nβ Key techniques: Propagate the ring image undistorted Detect the photons with a high efficiency (~20 hits/track) and with an excellent time resolution (<50 ps) Only MCP-PMTs can meet the requirements.
400 mm MCP-PMT for the TOP counter 4 Square shape multi-anode MCP-PMT with a large photocoverage Developed for the Belle II TOP counter at Nagoya in collaboration with Hamamatsu e Photon (Cross-section) Photocathode (NaKSbCs) MCP 2 4 4 anodes Oscilloscope (2.5 GHz bandwidth) 0.8 ns 13 10 mm 5.275 mm Micro channel e ~1 kv / 400 mm Small transit time spread (TTS) 50 mv @ 2.6 10 6 gain Output for single photons KT0552 ch5 2343 V The best time resolution (s~30 ps) of photon sensors
Performance of the MCP-PMT ADC distribution for single photons TDC distribution for single photons from picosecond pulse laser JT0886 ch5 QE distr. at 360 nm 5 KT0552 ch5 2140 V Photocathode MCP1 MCP2 Photocathode Typical QE spectrum Gain TTS s of 1 st Gaussian mean of the distribution = 41.8 ps = 5.1 10 5 (incl. ~17 ps laser pulse width and ~24 ps electronics jitter) KT0525
Performance in B-field 6 JT0743, ch6, 3240 V JT0743, ch6, 3240 V Nearly constant e B E 10 mm Shorter path in a higher B-field More bounces / Lower energy Higher gain / Lower secondary electron yield Less divergence in a higher B-field Rise and fall of the hole coverage or gain MCP1 MCP2
Number of PMTs Installed Mass-production of the MCP-PMTs Unprecedented production of 512 (and spare) MCP-PMTs. In parallel, R&D for life extension. Eventually three types of MCP-PMTs (Next talk by Muroyama-san) 700 600 500 400 300 200 100 0 ALD: Atomic Layer Deposition Conventional MCP (short lifetime) ALD MCP R&D Life-extended ALD MCP For replacement 2011 2012 2013 2014 2015 2016 2017 2018 2019 Succeeded in time for the TOP installation in May 2016. Mass-production is continued for the replacement of the 224 conventional MCP-PMTs in 2020 summer. 1st beam operation 7
Performance check at Nagoya The performance of every MCP-PMT was checked in automated test benches in a systematic way. 8 QE at peak (~360 nm) Collection efficiency TTS 29.3% avg. Total Conventional ALD Life-extended ALD 34.3 ps avg. Requirement: 24% min and 28% avg. (same gain of 2 x 10 6 ) JT0763_20140626 ch6 3460 V KT0162_20140612 ch6 2550 V Requirement: less than 50 ps The difference only at the tail, where the recoil photo electrons contribute, makes the difference of CE.
Performance check in 1.5 T The performance of every MCP-PMT was checked in a large dipole magnet at KEK. Checked the difference between 0 and 1.5 T. 9 gain(1.5 T) / gain(0 T) CE(1.5 T) / CE(0 T) TTS(1.5 T) TTS(0 T) Total Conventional ALD Life-extended ALD
PMT module assembly / installation 4 MCP-PMTs are assembled in a module. PMT window is glued on a wavelength filter, which cuts l 340 nm to suppress chromatic dispersion. Bubble free optical contact between the PMT module and the prism by a soft cast silicone cookie. 2.7 GSampling/s of PMT signal by switched-capacitor array ASIC (IRSX). [arxiv:1804.10782] Laser single photons for the in-situ calibration. 10 Pogo pin Readout electronics PMT module Silicone cookie CCD cameras LEDs / laser fibers Laser Prism Bar
number of channels ADC Threshold efficiency The gain of every MCP-PMT was adjusted to 5 x 10 5. Lower gain longer lifetime but lower threshold efficiency Evaluated the efficiency with single photons from the laser. 11 Laser signal Threshold Sampling time (ns) Data taken without discrimination Pedestal Signal 2 0 2 4 6 8 10 Output charge (10 5 e 0 ) f x = p 0 Τ x x 0 p 1 exp xτ p x 2 0 70 60 50 40 30 20 10 0 0.95 0.96 0.97 0.98 0.99 1 threshold efficiency
Beam operation MCP-PMT HVs were turned on during luminosity runs in Apr-Jul 2018. TOP counter worked for particle identification. First collision event on April 26 12 TOP detected Cherenkov photons Example of Cherenkov ring image
Beam background PMT hits are dominated by g rays from the accelerator g Compton scattering / pair creation in the quartz bar electrons Cherenkov photons MC estimation: 5-8 MHz/PMT at the design luminosity ~0.5 MHz/PMT in the start-up luminosity runs in 2018 Much higher than predicted, but still tolerable. 13 Accumulated output charge of each MCP-PMT Preliminary Kept below 0.023 C/cm 2 cf. QE drops by 20% at 0.3-1.7 C/cm 2 for the conventional MCP-PMTs
Evaluation of number of hits 14 Number of hits of Cherenkov photons for di-muon events MC based on the measured parameters of each component Quartz internal reflectance and transmittance MCP-PMT QE and collection efficiency (dark noise negligible) Readout efficiency (~77%, to be improved) and noise hits (a few %) Beam background hits (~1 hits/slot) Slot01 (life-extended ALD MCP-PMTs) Preliminary Slot11 (conventional MCP-PMTs) Preliminary Data (mean = 29.8) MC (mean = 26.1) Data (mean = 20.7) MC (mean = 20.2) The difference is under investigation.
Summary The MCP-PMT is one of the key components which bring the Belle II TOP counter into life. Succeeded in developing and producing 512 (and spare) MCP-PMTs for the Belle II TOP counter. ~34 ps TTS for every PMT 29.3% avg. QE at ~360 nm Work in 1.5 T Installation of the TOP counter finished in May 2016. The MCP-PMTs worked as expected in the first beam operation in Apr-Jul 2018. 15
QE measurement setup Measure the photocathode current with a picoammeter: QE MCP = I MCP IPD QE PD A 200 V 17 Photocathode MCP1 MCP2 MCP-PMT Light spot < 1 mm f Photodiode Slit Variable ND filter Sharp cut filters Monochromator Xe lamp Moving stage
Laser measurement setup Single photon irradiation to each channel one by one. 18 Dark box Reference PMT Moving stage ND filters MCP-PMT Fiber Slit Slit Light spot 1 mm f Laser MCP-PMT Pico-second pulse laser (l = 400 nm) Variable amp ATT Amp 10 db +33 db Discriminator Threshold: 20 mv ADC TDC +19.5~35 db
Installation of the TOP counter 19 ALD Life-extended ALD Conventional Installation of 16 TOP modules finished in May 2016. Viewed from the backward to the forward